Do you regularly add alt texts to your images?
Please be honest, I want to use this result in an essay.

​​ Boosts appreciated!

@ActuallyAubrey @finn : absolutely. When I'm low on energy or in pain, I just don't toot pictures. Which, honestly, is not that bad.
Also I avoid using gifs when I'm on my phone because I find them harder to ALT text, and I don't have the patience on a virtual keyboard.

@ladyteruki What do you think of #Alt4Me Hashtags or the like? I. e. the poster asking in such circumstances for others the write an alt-text, which the poster then edits to the formerly alt-text-free original post?
@ActuallyAubrey @finn

@ojelabii @ActuallyAubrey @finn : I sometimes do it for others but I prefer to avoid it. If a picture is communication, then surely ALT text is communication too. Being able to emphasize and describe what I want to discuss by using the image, is a way of connecting with the people who rely on ALT text, and more direct than playing telephone with an intermediary.
That said, in the abstract it's not a bad thing in general (and way better than AI description). I just like it less personally.

@ladyteruki I understand. Thanks!
@ActuallyAubrey @finn

@finn@ice.finnley.dev I choose no 2, use @altbot@fuzzies.wtf if the alt text is insufficient.

@ahrienby@transfem.social @finn@ice.finnley.dev there are multiple problems with using a bot to generate alt text as your main option.

First, it's the technological reality that a bot might not recognize stuff in the picture (there was one a while ago that Jolene used to have fun with, where the bot went on a loop and blew itself up), or might miscategorize what it actually sees

Second, the bot doesn't know the intention. If it was made by whoever is posting, they can explain some parts of the image that could be misunderstood (just the other day the foxes in love alt text explained that one of them wanted to eat a pumpkin whole, while I thought it was just carrying it with its mouth)

Third, even if neither of the previous 2 apply, sometimes the order in which you describe the image can have the same effect of delaying the punchline like when you tell a joke, or how you'd see a visual pun. A sighted user can look back and forth in the image easily, a screen reader has to reread the entirety of the alt text every time, so if the joke was too early, it can be missed without a reread.

Fourth, it's not always obvious to a bot what context is well known and what needs to be explained. Heck even for a human, I don't always know when to describe a meme and when to use it's name, let alone if I feel like I need to explain that this is a callback to an in-joke in a circle that the reader might not be a part of.

I'm purposefully leaving out questions about morality and stuff (altbot is an LLM) because for one, there are people who will never be convinced and so having the option is still good because accessibility is a net good, and for another this will derail the conversation about why a bot backup is good, but it should remain a backup.